LinkedIn recommendations are one of the most under-used assets on a candidate's profile. Most people have zero or two, both generic, written by the wrong people. The right approach produces a small set of specific, recent recommendations from people who actually worked with you — and recruiters do read them, sometimes.

This post is the version of recommendation-collecting that produces real recommendations from people who write them.

Why this is worth doing at all

Recruiter behavior surveys consistently show that 35-45% of recruiters read at least one LinkedIn recommendation during the initial-screen process. Recommendations aren't a top-three factor in hiring decisions, but they're meaningfully more read than candidates assume.

The recruiter behavior is usually: scan for recommendations after looking at experience and education, read the most recent one or two, glance at who wrote them. The signal recruiters extract is specific. A generic recommendation ("Jane was great to work with — she's smart, hardworking, and a pleasure as a colleague") provides almost no signal. A specific recommendation ("Jane led the platform migration in 2024 and built the team's first observability dashboard") provides real evidence that supports the claims on the resume.

Two implications:

• A small number of specific recommendations beats a larger number of generic ones. Three specific recommendations from people who worked with you closely are worth more than eight vague ones from acquaintances.
• The recommendation is a credibility multiplier on what's already there. It doesn't independently make a hire decision. It backs up the resume claims with a second voice.

Who to ask

The right recommenders are the people who saw your work closely enough to write something specific. The wrong recommenders are people who like you but don't have the texture.

Ask:

• Your direct manager or skip-level, especially if you had a defined working relationship over at least six months.
• A peer you collaborated with closely on a specific project. Peer recommendations are often the most vivid because peers see day-to-day work that managers don't.
• A client or stakeholder you delivered for, whose project went well. External recommendations carry weight because they're harder to manufacture.
• A direct report you managed for at least six months. Reports who choose to recommend their managers tend to write specific, credible recommendations.
• A cross-functional partner who can speak to a specific deliverable. The product manager you worked with closely. The lead from another team who joined your initiative.

Skip:

• Senior executives you met in passing. The recommendation is generic by necessity, and recruiters can tell.
• Anyone you haven't spoken to in 3+ years. They can't write specifically because they don't remember specifically.
• Friends or family members. These are immediately readable as personal.
• Five peers at the same level from the same team. This pattern looks staged, even when it isn't.
• Your current manager, in most cases. Asking creates awkward dynamics — they may worry you're job-searching, and the recommendation often becomes generic to avoid commitment.

The strongest pattern is two to four recommendations: one from a former manager, one from a peer, one from a client or stakeholder, optionally one from a direct report. That mix covers the angles a recruiter is reading for.

The five-move ask

A working ask has five parts:

Pick the right person, not the most senior. Specificity beats title. A peer recommendation that names specific projects beats a VP recommendation that says "always impressed by Jane's contributions" almost every time.

Wait for the right moment in the relationship. Ask after a project concluded well, after you delivered something they noticed, or while you're still in regular contact. Asking 18 months after you stopped working together makes the ask harder for the recommender — they have to reconstruct memory, which usually produces either a generic recommendation or a polite delay that turns into a no.

Make the ask outside the LinkedIn "Ask for a recommendation" button. The in-app request generates a template message that recommenders deprioritize. A direct DM or email ("Hey — I'm refreshing my LinkedIn and would love a recommendation from you if you'd be up for writing one") gets much higher response rates. The request feels personal because it is personal.

Make it easy: suggest the angle and the length. This is the move that doubles response rates. Recommenders aren't lazy; they're busy, and "write something nice about me" is a harder request than "could you write 3-4 sentences about the work we did on the X migration, focusing on the database side?" The angle gives them a starting point. The length gives them permission to be brief.

Follow up once, then stop. If they haven't written within two weeks, send one polite nudge. If still nothing, drop it — and don't bring it up the next time you interact. Half of agreed-to recommendations never get written. This is normal, not personal.

What to write in the actual request

A working request, in DM or email:

Hey [Name],

I'm doing a quiet refresh of my LinkedIn — would you be open to writing me a short recommendation? No pressure at all if it's not the right moment.

If it helps, the project I'd most love it to reflect is the [specific project] we worked on, where I [specific contribution]. 3-4 sentences is more than enough — I know you're busy.

Happy to write you one in return, of course, if I haven't already and there's an angle that would be useful for you.

The reciprocal offer at the end is genuine — recommendations from people who worked with you closely are the easiest ones for you to write, and they're worth the 15 minutes.

What to do when you receive one

When the recommendation arrives, two moves:

Don't edit it heavily. LinkedIn lets you request changes, but doing so substantively often offends the recommender. If there's a factual error (wrong date, wrong company name), ask politely. If you just wish it were warmer, leave it.

Send a real thank-you, not just a LinkedIn auto-message. A short DM or email thanking them specifically — "the way you described the migration captured exactly what I was hoping someone would notice" — closes the loop and keeps the relationship warm.

Reciprocity without staging

LinkedIn shows when two people have recommended each other. A reciprocal recommendation isn't damning — most close working relationships produce mutual recommendations naturally — but a wall of obviously-staged "Jane wrote one for me on Tuesday and I wrote one for Jane on Wednesday" recommendations reads as theatrical. The fix is to write the reciprocal ones on a delay, when there's a real angle, not as immediate quid-pro-quo.

For the broader relationship between LinkedIn and your resume as two distinct documents that should reinforce rather than duplicate each other, see linkedin-vs-resume. For the related question of how to handle other LinkedIn social proof like skills endorsements, see linkedin-skills-endorsements-myth.

What this isn't

A few clarifications:

• It's not a guarantee of more interview callbacks. Recommendations help modestly. They're a credibility multiplier, not a primary signal.
• It's not a license to ghostwrite your own recommendation. Some recommenders will ask you to draft something. Politely decline. The whole point is that someone else wrote it.
• It's not equally valuable in every field. In sales, finance, and consulting — where individual performance is verifiable and reference-checked — recommendations matter more. In commodity-skill roles, they matter less.

The short version: ask two to four specific people, by direct message, with a suggested angle and length, after a project went well. Follow up once. Specificity beats title; recency beats prestige. Half of recommendations never get written, and that's fine.